This episode of More or Less was broadcast on Thursday, 26 June, 2003 on BBC Radio 4 at 15:00 BST.

When it comes to earnings, where is the middle? According to Peter Hain, the cabinet minister who caused such agitation last week when he said there should be a widespread debate about tax levels, an increasing number of middle income earners are now paying top rate tax.

His remarks were widely interpreted as an argument for higher tax on the very rich in order to take those "middle income" earners out of the higher tax bracket. That puts his sense of the middle at roughly �35,000. Is he right?

His remarks illustrate the importance of knowing the numbers when making arguments of principle. Yet much media coverage managed to report the story with scarcely a mention of the relevant figures.

More or Less this week found out how well informed we are about who earns what and who pays what in tax, and asks how the debate might change if the numbers were more widely known.

You can use our table to see how your salary compares to the rest of the UK population. Look up your take home pay (after tax and national insurance) of your household in pounds per week in the figures on the left (so if there are two of you earning, add the wages together).

Then find out which percentage group you're in on the right, ie 0-5% means you're in the bottom 5% and 95-100% means you're in the top 5%. The top half starts at �320 per week.

And we also looked at gambling - where the numbers seem to make little difference to people's willingness to chance their money.

Also in the programme, the importance of a round number. How many times do you see the claim that "1 in 4" of the population suffers from X? How come so many social causes or campaigns come up with a "1 in 4" statistic?

Cancer risk statistics

Following our story last week about the way cancer risks are often presented in ways that can make the danger seem greater than it is, Dr Jane Hutton of Warwick University suggested this code of conduct for the public presentation of medical statistics:

1. The baseline risk, in a defined population and a defined period, must be stated.

2. The risk (population, period) allowing for whatever factor (e.g. alcohol) should then be given.

3. A statement of the risks in terms of a thousand or a hundred people should be given.

4. A margin of error, or other indication of uncertainty, should be given.

5. If the risk relates to increasing levels of a factor, (e.g. the number of drinks per day), the risk should be given for a low and a high level of the factor (e.g. 1 drink per day, 5 drinks per day).

We would welcome your views.
Click here to have your say